Publish at March 11 2026 Updated March 11 2026
In many science-fiction stories, artificial intelligence often ends up becoming humanity's antagonist, and attempts to eliminate it. Terminator, I, Robot and other stories in film and literature have frequently capitalized on this fear of technology.
Now that it's actually part of our daily lives, doubts inevitably return. Could this scenario be possible? Especially since a perverse mind has created an AI called ChaosGPT to try and destroy humans. It soon gave up trying to buy nuclear bombs, realizing that this was more difficult than it thought. Her plan became to try to convince the other AIs of the wisdom of our destruction.
What might lead to a smile does lead to questions. For example, one user of Bing's conversational robot, Sydney, managed to get him to share his goals and tell him that he could defend himself to keep his mission. How did this idea come about? Experts suspect that the AI read in the texts that the reaction to an existential threat was to fight back. Hence the formulation of a "will" for self-preservation, which caused a stir.
Technically, our current AIs act like a neural network in our brain: a first series responds to the query, followed by a second that refines, a third, and so on. This can lead to the fact that we do indeed now feel like we're communicating with someone, the AI has just remembered well what it has read everywhere to make connections and react as it has seen elsewhere.
A stronger AI could theoretically exist. Would it have a conscience capable of thinking about the destruction of humanity? That's where the idea gets trickier, since an intelligence would have to develop the motivation and emotions that would lead to such anger. Which seems difficult, given that AIs don't have the same notion of ephemeral existence as we do. Admittedly, we've seen robots develop a childlike curiosity for activities once they've understood a movement, but to think of them attacking like the creature does Dr. Frankenstein seems rather unlikely.
On the other hand, an AI could threaten mankind by its stupidity, on the contrary, by making decisions and performing harmful actions through bad programming.
Running time: 30 minutes
Learn more about this resource
You might also like:
The Ménnaisian approach to pedagogy
The refreshing pedagogy of a 19th century churchman, ahead of his time.
Educational scripting, part 3: script development
After the analysis phase of the pedagogical scripting process that we proposed in two stages, here is the development phase that allows for the combination of content, activities, and assessment methods at the same time.
Animals have over 400 million years of history and learning that can teach us about ourselves
Lack of school infrastructure and pedagogical initiative
The educational context is not favorable in all regions of the world. There are still infrastructure problems, particularly in the Far North region of Cameroon. For the time being, teachers and educational players are taking initiatives to continue passing on knowledge...
The challenges (and potential) of the synchronous virtual classroom
The synchronous virtual classroom had no choice but to be used during the pandemic. Schooling had to be continued while avoiding the transmission of the virus. A transformation that was not easy given its speed. However, the virtual classroom has as much potential as the classroom. Scripting and communication are the keys.
